Henning Schoenenberger, SpringerNature, GER
Mads Rydahl, Unsilo.com, DK
Ashleigh Gardner, Wattpad, CAN
Uli Klopotek, Klopotek, GER
Henning Schoenenberger, SpringerNature, stressed that managing (and reducing) the “information overload” in scientific publishing is crucial. He told the audience that they put themseves to he challenge of publishing the first machine learning-generated scientific book “in full disclosure”, providing full transparency. It was generated using an algorith on the basis of a re-combined accumulation of content in a given subject area – on lithium-ion batteries. Henning said that this worked out quite well; they’re planning to do this in other fields as well. (The ebook is available at SpringerLink for free download.) – He said that for such publications, humans will in the future have an interface to influence the creation of the machine-generated book, such as number of terms/keywords in a chapter, (confirmed/curated) relevance of selected content, etc.
This, Henning stressed, is not a just a problem of technology, but also of publishing, in order to tackle the information overload. “The main use case is about reducting the mass of content in scholarly communication and the redundancy of already known information in research.”
Mads Rydahl said that “we define meaning from intent”. “AI has always been pre-programmed intent, but so does biology”. He believes that “intent” is the keyword for the entire discussion, and that instead of using the phrase “artificial intelligence”, we should talk about “smart products”, to recognise the human intelligence required for good product design.
At #pf19, @rydahl just gave a great presentation about #AI in #publishing, recommender systems, other cool stuff they do @unsiloproduct and why „smart“ is a mich better word than „intelligent“ when it comes to products. Thanks, Mads!
— robyn kerkhof (@thefoxylinguist) 10. Mai 2019
Uli Klopotek asked what the role of the publisher is when developing content? What is the outcome if not the best content is being used? This applies, he said, to Wattpad as well as to SpringerNature.
Moderator Porter Anderson said that this would be “garbage in, garbage out”.
Henning Schoenenberger said that they only use peer-reviewed content for their related projects.
Ashleigh Gardner stressed that they use different kinds of models for selecting stories for Wattpad Books. It’s not that easy to determine what is “good”, she explained.
Uli Klopotek thinks that many publishers nowadays don’t understand some fundamental changes in the market, such as using AI. Many products are not successful, and the time to market is to long. “Always more and more is not better” (on annual publishing programs). Instead, getting to really know your target groups would be relevant (modern technology could help to overcome traditional program planning). “Do you really know what your readers actually want?”
This is especially important in times of crisis, Porter Anderson said, which has been confirmed by Jesús Badenes yesterday. In these times, you cannot waste money on titles you know (because readers tell you) will not work.
Uli Klopotek also referred to the 6 m “lost readers” – he said that publishers do not make enough effort to win them back to be readers by providing different types of stories/books/products.
Mads Rydal mentioned the topic of curation which “is a burden for many publishers”. He said that in the music business, we’re witnessing a change, users can now be curators (on platforms such as Spotify). Wattpad, of course, “is a world of users”, Porter said. He also said that “Wattpad is a fascinating shift, but also a scary shift” for publishing.